Gemini 2.5 Flash (cloud fallback)

A high-performance redundancy tier that bridges on-device inference with Gemini 2.5 Flash to ensure consistent AI availability across all hardware.

Gemini 2.5 Flash (cloud fallback) acts as the high-capacity redundancy layer for hybrid AI deployments (typically via the Firebase AI Logic SDK). This architecture prioritizes on-device models like Gemini Nano for privacy but triggers a seamless cloud transition when tasks demand the model's 1 million token context window or advanced thinking capabilities. It maintains sub-second response times for high-volume tasks like real-time summarization and structured data extraction. By using this fallback, developers ensure 100% feature parity across diverse hardware profiles without manual routing logic.

https://firebase.google.com/docs/ai/hybrid-inference